A neural network trained by PitchBook concluded that much of what PE firms call "operational alpha" can be replicated by systematically selecting companies in favorable sectors (like tech), applying more leverage than public market counterparts, and benefiting from market-wide multiple expansion, rather than superior operational improvements.
PitchBook's analysis of "marquee" or household-name buyout managers shows a clear downward trend in performance. On a capital-weighted basis, these large funds have seen their relative performance scores degrade over time, recently falling below the median and underperforming the rest of the fund universe.
The decision to allocate to massive, well-known PE funds is often driven by behavioral factors like career risk. Similar to the old adage "you don't get fired for buying IBM," allocators choose brand-name managers for defensibility and convenience, even if performance data suggests smaller funds may offer better returns.
The long-held belief in performance persistence in private equity is weakening. Early academic studies were flawed because they analyzed finalized fund data, whereas re-up decisions are made mid-fund life. Newer data shows that picking managers based on their last fund's performance is "certainly not foolproof, if helpful at all."
Despite massive capital flows to the largest funds, middle-market funds have demonstrated superior performance for over ten years post-GFC. The challenge for large LPs is not performance, but the structural difficulty of deploying large checks into smaller funds without becoming over-diversified and reverting to the mean.
PE returns appear artificially smooth because they are based on infrequent, private valuations. This "volatility laundering" mathematically lowers standard deviation and correlation inputs in asset allocation models. It makes the asset class seem less risky, causing models to systematically over-allocate capital to it based on flawed risk assumptions.
Data from 2000-2021 shows a startling trend: in 9 of those 21 vintage years, the largest buyout deals experienced EBITDA margin declines post-acquisition. This contradicts the core private equity value proposition of improving operational efficiency and suggests returns are heavily reliant on financial engineering rather than making businesses fundamentally better.
A huge LP like CalPERS may need to deploy $15-20 billion annually. This operational constraint makes allocating to smaller, better-performing middle-market funds impractical. The need to write huge checks forces them into mega-funds that can absorb the capital, creating a demand-driven consolidation cycle independent of performance.
The success of the Yale/Swenson model was predicated on a first-mover advantage in an uncrowded market. Today, with trillions in AUM, the private equity landscape is intensely competitive. Increased competition among LPs for top managers and among GPs for deals has eroded the pricing power and advantages that pioneering endowments once enjoyed.
A common practice in institutional portfolio modeling is to simply assume private equity will outperform public equity by a fixed premium, often 300 basis points. This simplistic assumption, rather than rigorous analysis, drives allocation decisions and creates self-fulfilling demand for the asset class, irrespective of actual, risk-adjusted performance.